building real-time analytics apps with hana
TRANSCRIPT
Building Real-Time
Analytics Apps with HANA
Why SAP HANA Now?
‘Large’ Data
Columnar Databases
Moore’s Law
Inflection Point?
What is SAP HANA?
An Appliance?
A Database / RDBMS?
A Platform?
Answer – All of the Above
SAP HANATM - Preconfigured Appliance
Software + Hardware
IBM and others…
In-Memory Computing Engine Software
Data Modeling and Data Management
Real-time Data Replication Data Services for SAP
SAP Business Suite, SAP BW and 3rd Party Systems
Capabilities Enabled
Analyze information in real-time at unprecedented speeds on large volumes of non-aggregated data
Create flexible analytic models based on real-time and historic business data
Foundation for new category of applications (e.g., planning, simulation) to significantly outperform current
applications in category
Minimizes data duplication
12
In-Memory
Row and Column Store Compression Partitioning
In-Database Computing No Aggregate Tables Real-Time Replication
Insert Only on Delta
Column = Fast Queries
5 – 20x
Analyze Large Data Sets
Complex Computations
Flexible Modeling
No Data Duplication
Fast Data Loads
How does it work
Getting Stuff into HANA
Sybase Replication Server (SRS)
• Database Table Replication i.e. near real-time (10 secs)
• Originally envisaged to be dominant
• DB2 Unicode Only
• No Compressed Tables
• No license support for Oracle/MSSQL
• Support for 1 ERP
Data Services 4.0 (DS)
• Extract, Transform & Load (ETL)
• Allows use of ERP Business Content Extractors
• Non-real-time (10 mins)
• Almost any data source
System Landscape Transformation (SLT)
• Technology from Test Data Migration Server (TDMS)
• Application-server trigger-based replication
• Avoids Oracle/MSSQL license problems
• Near real-time
SLT – A Bit More Detail
or
Getting Stuff Out of HANA
Today - SAP HANA 1.0 SP02
• Financial Services
Hedge fund trading analysis
Real-time systematic risk management and reporting based on market trading exposure
• Consumer Products Goods
Supplier Risk Management
Inventory Recall Reconciliation Product trace
Product Lifecycle and Cost management
EPA standards compliance
On-Shelf Availability Analysis
Real-time Warranty and defect analysis
• Manufacturing
Manufacturing Production and Maintenance
Analytics
Real-time Asset Utilization Analytics
• Oil & Gas
Operations Performance Management
Business Transformation
Oil Well Performance Analysis
Supply Chain Network Visibility
Spend Performance Analysis
• Retail
Store Operations
Point-of-Sales Retail Sales Analytics
Load Balance Optimization
Real-time completive multi-channel pricing
Product damage, returns and spoilage
These are All Custom-Built Scenarios and require a business case
Today - SAP HANA RDS
• RDS exists to replicate CO-PA to HANA
• Includes Data Replication (SLT)
Transformations (SAP HANA Studio)
Sample Reports (BI4)
Building a Business Case
HANA represents a significant investment Appliance:
Hardware
Software Licenses
Services Installation
Configuration: SAP or Partner Services
SAP & Partners provide Value Engineering Services Will a point solution provide ‘Proof of Value’?
Where is the process benefit?
Can business process change?
Who is impacted?
SAP HANA Demo
SAP HANA Demo
SAP HANA Studio
Modeling: SAP HANA Studio:
Creating Attribute View
Modeling: SAP HANA Studio:
Defining Attributes
Modeling: SAP HANA Studio:
Data Preview
Modeling: SAP HANA Studio:
Creating Hierarchies
Modeling: SAP HANA Studio:
Creating Analytic View
SAP HANA Studio:
Administration Console
Questions?
For more information
IBM-SAP Alliance Web Site www.ibm-sap.com/hana
Partnerworld
www.ibm.com/partnerworld/sap
Bluefin Web Site www.bluefinsolutions.com
Bluefin Insights
www.bluefinsolutions.com/insights/
Thank You
Appendices
• SAP HANA 1.0 SP03 (BW on HANA)
• SAP HANA Apps
• SAP HANA Beyond 2012
• Scale Out for SAP HANA
Appendix 1
SAP HANA 1.0 SP03
BW on HANA
SAP HANA 1.0 SP03
(AKA BW on HANA)
Appendix 2
SAP HANA Apps
H1 2012: SAP HANA Apps
• Strategic workforce planning (out now)
• Sales and operations planning (due Q3)
• Cash and liquidity management (due Q3)
• Trade promo management (due Q4)
• Smart meter analysis (due Q4)
• Profitability engine (due Q4)
• Customer revenue performance management (due Q4)
• Merchandising and assortment management (due Q4)
• Energy management for utility customers (due Q4)
• Customer-specific pricing (due Q4)
• Intelligent Payment Broker (Q4)
• SAP traditionally sells to industry
• SAP HANA 1.0 is a platform
• SAP HANA Apps are business content built on SAP HANA 1.0
SAP HANA Apps: SWP
Appendix 3
SAP HANA Beyond 2012
SAP HANA: Beyond 2012
Platform Consolidation
All applications (ERP and BW) run on data residing in-memory
Analytics and operations work on data in real time
In-memory computing executes all transactions, transformations, and complex data processing
Real Time Data Feed (2012/2013?)
Applications write data simultaneously to traditional databases as well as the in-memory computing
Appendix 4
Scale Out for SAP HANA
(AKA BW on HANA)
node01
GPFS
•replication server •SAP HANA studio
data01 log01
- SAP HANA DB
master node
DB partition 1
HDD SSD
SAP HANA DB
Primary dataset data01 log01
Currently only single server
configurations available
What about
growth and HA ?
Scale out for SAP HANA
node01
GPFS
node02
•replication server •SAP HANA studio
- SAP HANA DB
backup master
GPFS
HDD SSD
data01 log01
- SAP HANA DB
master node
DB partition 1
HDD SSD
SAP HANA DB SAP HANA DB
1st replica
Primary dataset
DB partition 2
IBM configurations delivered with GPFS
shared file system for seamless growth
Scale out for SAP HANA
node02 node01
GPFS
node03
•replication server •SAP HANA studio
- SAP HANA DB
GPFS
HDD SSD
data01 log01
- SAP HANA DB
master node
DB partition 1
HDD SSD
SAP HANA DB SAP HANA DB
DB partition 2 DB partition 3
- SAP HANA DB
backup master
HDD SSD
Primary dataset
IBM configurations delivered with GPFS
shared file system for seamless growth
Currently 4 node
cluster certified
( 4x 256GB &
4x 512GB )
1st replica
Scale out for SAP HANA
node03 node02 node01
GPFS
node04
•replication server •SAP HANA studio
- SAP HANA DB
standy node
GPFS
HDD SSD
data01 log01
- SAP HANA DB
master node
DB partition 1
HDD SSD
SAP HANA DB SAP HANA DB
DB partition 2 DB partition
- SAP HANA DB
backup master
HDD SSD
- SAP HANA DB
DB partition 3
HDD SSD
1st replica
Primary dataset
x
Introduce Standy nodes into cluster – initially
no specific identity and no persistent data
Outlook – HA for SAP HANA
(under validation)
SAP HANA DB handles the failover to the hot-
standby node in case of a node failure.
node03 node02 node01
GPFS
node04
•replication server •SAP HANA studio
- SAP HANA DB
GPFS
HDD SSD
data01 log01
- SAP HANA DB
master node
DB partition 1
HDD SSD
SAP HANA DB SAP HANA DB
DB partition 2 DB partition
- SAP HANA DB
backup master
HDD SSD
- SAP HANA DB
HDD SSD
1st replica
primary dataset
x DB partition 3
Apply
Re-Do Log
Recover DB
partition 3 data
Outlook – HA for SAP HANA
(under validation)